Atmos. Chem. Phys., 21, 14333–14349, 2021https://doi.org/10.5194/acp-21-14333-2021© Author(s) 2021. This work is distributed underthe Creative Commons Attribution 4.0 License.

Impact of pyruvic acid photolysis on acetaldehydeand peroxy radical formation in the boreal forest:theoretical calculations and model resultsPhilipp G. Eger1, Luc Vereecken2, Rolf Sander1, Jan Schuladen1, Nicolas Sobanski1, Horst Fischer1, Einar Karu1,Jonathan Williams1, Ville Vakkari3,4, Tuukka Petäjä5, Jos Lelieveld1, Andrea Pozzer1, and John N. Crowley1

1Atmospheric Chemistry Department, Max-Planck-Institute for Chemistry, 55128 Mainz, Germany2Institute for Energy and Climate Research: IEK-8, Forschungszentrum Juelich, 52425 Juelich, Germany3Atmospheric Composition Unit, Finnish Meteorological Institute, 00101 Helsinki, Finland4Atmospheric Chemistry Research Group, Chemical Resource Beneficiation,North-West University, Potchefstroom, South Africa5Institute for Atmospheric and Earth System Research (INAR)/Physics,Faculty of Science, University of Helsinki, Helsinki, Finland

Correspondence: John N. Crowley (john.crowley@mpic.de)

Received: 18 September 2020 – Discussion started: 20 October 2020Revised: 1 September 2021 – Accepted: 7 September 2021 – Published: 28 September 2021

Abstract. Based on the first measurements of gas-phasepyruvic acid (CH3C(O)C(O)OH) in the boreal forest, we de-rive effective emission rates of pyruvic acid and comparethem with monoterpene emission rates over the diel cycle.Using a data-constrained box model, we determine the im-pact of pyruvic acid photolysis on the formation of acetalde-hyde (CH3CHO) and the peroxy radicals CH3C(O)O2 andHO2 during an autumn campaign in the boreal forest.

The results are dependent on the quantum yield (ϕ) andmechanism of the photodissociation of pyruvic acid andthe fate of a likely major product, methylhydroxy carbene(CH3COH). With the box model, we investigate two differ-ent scenarios in which we follow the present IUPAC (IU-PAC Task Group on Atmospheric Chemical Kinetic DataEvaluation, 2021) recommendations with ϕ= 0.2 (at 1 bar ofair), and the main photolysis products (60 %) are acetalde-hyde+CO2 with 35 % C–C bond fission to form HOCO andCH3CO (scenario A). In the second scenario (B), the forma-tion of vibrationally hot CH3COH (and CO2) represents themain dissociation pathway at longer wavelengths (∼ 75 %)with a ∼ 25 % contribution from C–C bond fission to formHOCO and CH3CO (at shorter wavelengths). In scenario 2we vary ϕ between 0.2 and 1 and, based on the results ofour theoretical calculations, allow the thermalized CH3COH

to react with O2 (forming peroxy radicals) and to undergoacid-catalysed isomerization to CH3CHO.

When constraining the pyruvic acid to measured mixingratios and independent of the model scenario, we find thatthe photolysis of pyruvic acid is the dominant source ofCH3CHO with a contribution between ∼ 70 % and 90 % tothe total production rate. We find that the photolysis of pyru-vic acid is also a major source of the acetylperoxy radical,with contributions varying between ∼ 20 % and 60 % de-pendent on the choice of ϕ and the products formed. HO2production rates are also enhanced, mainly via the forma-tion of CH3O2. The elevated production rates of CH3C(O)O2and HO2 and concentration of CH3CHO result in significantincreases in the modelled mixing ratios of CH3C(O)OOH,CH3OOH, HCHO, and H2O2.

1 Introduction

Organic acids play a crucial role in tropospheric chemistry,impacting secondary organic aerosol formation, air qual-ity, and climate (Kanakidou et al., 2005; Hallquist et al.,2009). Pyruvic acid (CH3C(O)C(O)OH), an organic acidthat is central in plant metabolism as part of the Krebs cycle

Published by Copernicus Publications on behalf of the European Geosciences Union.

14334 P. G. Eger et al.: Impact of pyruvic acid photolysis on CH3CHO and RO2

Figure 1. Wavelength-resolved photolysis rates (Jpyr) for13 September 2016 at solar noon. Jpyr was calculated using a pho-tolysis quantum yield of 1 and the absorption cross sections at 298 Kpreferred by IUPAC Task Group on Atmospheric Chemical KineticData Evaluation (2021).

(Walker, 1962), is found in tropospheric air in the gas phaseas well as in the aerosol phase, especially in the boundarylayer over vegetated regions. Gas-phase mixing ratios rang-ing from a few to several hundred parts per trillion (pptv)have been reported in various locations around the world, in-cluding the tropical rainforest, boreal forest, rural areas withtemperate forest, and regions influenced by urban outflow. Arecent overview of existing measurements of gas-phase pyru-vic acid is given by Eger et al. (2020).

A known source of pyruvic acid is the photo-oxidation ofisoprene, via the ozonolysis of methyl vinyl ketone and sub-sequent hydrolysis of the Criegee intermediates (Jacob andWofsy, 1988; Grosjean et al., 1993; Paulot et al., 2009). Fur-ther potential sources are the photolysis of methylglyoxal(Raber and Moortgat, 1995), the gas-phase photo-oxidationof aromatics in the presence of NOx (Grosjean, 1984; Praplanet al., 2014), the aqueous-phase oxidation of methylglyoxal(Stefan and Bolton, 1999), and reactions taking place withinbiomass burning plumes (Andreae et al., 1987; Helas et al.,1992). In addition, pyruvic acid has been reported to be di-rectly emitted from vegetation (Talbot et al., 1990; Jardineet al., 2010a, b; Eger et al., 2020). Compared to acetic acid,the presence of a second (non-acidic) carbonyl group impartson pyruvic acid an absorption spectrum that extends from ul-traviolet to visible wavelengths (see Fig. 1), and photolysis isa major sink of pyruvic acid in the boundary layer, with de-position and heterogeneous uptake to the aerosol phase alsocontributing to its removal. Photolysis of pyruvic acid in airresults in a number of different radical and stable products;the major ones are expected to be acetaldehyde, HO2, andCH3C(O)O2 (more details are presented in Sect. 1.1). Theseproducts can have a significant impact on tropospheric chem-istry, e.g. via the formation of peroxyacetyl nitrate (PAN),peracetic acid, and formaldehyde (HCHO).

Global models have recently revealed discrepancies be-tween simulated and measured acetaldehyde concentrations(Millet et al., 2010; Wang et al., 2019; Wang et al., 2020).Wang et al. (2020) reported CH3CHO mixing ratios thatwere up to a factor of 10 higher than predicted by a globalchemistry-transport model (EMAC) in the marine bound-ary layer around the Arabian Peninsula, implying missingsources of CH3CHO in remote and polluted regions. Wanget al. (2019) also found that models systematically underesti-mate CH3CHO compared to observations, implying a miss-ing source of acetaldehyde in the remote troposphere. Thisfinding was supported by the simultaneous measurement ofperacetic acid (which is formed, for example, via the degra-dation of acetaldehyde in remote environments) with the or-ganic aerosol source of CH3CHO also being insufficient toexplain the results. Instead, Wang et al. (2019) suggested thatCH3CHO arises from the degradation of gas-phase organiccompounds. Pyruvic acid, among other organic acids in thegas and aerosol phase, might be one of the compounds thatcan be converted to acetaldehyde in the remote troposphere,and its integration into global models might contribute to re-solve discrepancies, especially in forested regions.

Generally, field measurements as well as modelling andlaboratory-based kinetic studies on pyruvic acid are limited,and its impact on atmospheric chemistry is still poorly under-stood. In this study we highlight the potential role of pyruvicacid in the boreal forest, one of the largest terrestrial biomeson Earth. For this, we use data from a measurement campaignin 2016 (IBAIRN, Influence of Biosphere–Atmosphere Inter-actions on the Reactive Nitrogen budget).

The photolysis of pyruvic acid: quantum yields andproducts

Because its reaction with OH is slow(kOH+pyruvic acid= 1.2× 10−13 cm3 molecule−1 s−1 at298 K; Mellouki and Mu, 2003), photolysis and dry depo-sition are the dominant loss terms for gas-phase pyruvicacid. Heterogeneous uptake to atmospheric aerosols is alsocalculated to be inefficient during the IBAIRN campaignin the boreal forest (see below), where particle surfacearea densities were of the order of 2× 10−7 cm2 cm−3, andthe particles contained a large organic fraction (Liebmannet al., 2019) that is likely to reduce the uptake coefficientcompared to that measured for pure aqueous particles(γ = 0.06; Eugene et al. 2018).

The photodissociation of pyruvic acid at actinic wave-lengths is not well understood. According to the most recentIUPAC Task Group on Atmospheric Chemical Kinetic DataEvaluation (2021) evaluation, which considers experimentaldata until 2017, the three thermodynamically accessible pho-tolysis channels are

Atmos. Chem. Phys., 21, 14333–14349, 2021 https://doi.org/10.5194/acp-21-14333-2021

P. G. Eger et al.: Impact of pyruvic acid photolysis on CH3CHO and RO2 14335

CH3C(O)C(O)OH+hν→ CH3CHO+CO2 (R1)→ CH3C(O)OH+CO (R2)→ CH3CO+HOCO. (R3)

IUPAC recommends a photodissociation quantum yield(ϕ) of 0.2 at 1 bar pressure (i.e. for boundary layer condi-tions), with branching ratios of 0.6, 0.05, and 0.35 for Re-actions (R1), (R2), and (R3), respectively, which implicitlyassumes that the initially formed carbene (CH3COH) imme-diately isomerizes to acetaldehyde. The radical products ofReaction (R3) (CH3CO and HOCO) react rapidly in air toform peroxy radicals (Reactions R4 and R5).

CH3CO+O2+M→ CH3C(O)O2+M (R4)HOCO+O2→ HO2+CO2 (R5)

The formation of methylhydroxy carbene (CH3COH) asan intermediate in pyruvic acid photolysis has been postu-lated for many years (Vesley and Leermakers, 1964; Ya-mamoto and Back, 1985). Schreiner et al. (2011) observedisomerization of singlet CH3COH to acetaldehyde in an Armatrix at 11 K; their high-level theoretical analysis revealedhigh barriers for isomerization, where H-atom tunnellingthrough the energy barrier led to a lifetime of about 1 h at11 K, which favoured the formation of acetaldehyde over thatof vinyl alcohol. Only very recently has CH3COH been de-tected experimentally as a product of pyruvic acid photolysisin the gas-phase (Samanta et al., 2021) and its unimolecu-lar isomerization to both CH3CHO and CH2=CHOH con-firmed to be efficient at the experimental pressure of a fewmillibars (mbar) of helium. Indeed, Samanta et al. (2021)show that, at a photolysis wavelength of 351 nm (close tothe maximum cross-section of pyruvic acid) formation ofan energy-rich carbene (CH3COH#) and CO2 (Reaction R6)is essentially the only product channel operating. CH3CHOand CH2=CHOH were formed subsequently (at ∼ 2 : 1 ra-tio favouring CH3CHO) in the unimolecular isomerizationof CH3COH# (Reactions R7 and R8).

CH3C(O)C(O)OH+hν→ CH3COH#+CO2 (R6)

CH3COH#→ CH3CHO (R7)

→ CH2=CHOH (R8)

Samanta et al. (2021) suggest that, at ambient pressures, asignificant fraction of the energized, nascent carbene will bedeactivated by collisions with air (Reaction R9) and the ther-malized carbene (CH3COH), which can no longer rapidlyovercome the barriers to isomerization, may react with oxy-gen or water vapour (Reed Harris et al., 2016, 2017b; Egeret al., 2020; Samanta et al., 2021) (Reactions R10 and R11).

CH3COH#+M→ CH3COH+M (R9)

CH3COH+O2→ CH3C(O)+HO2 (R10)CH3COH+H2O→ CH3CH(OH)2 (R11)

As summarized by IUPAC, there have been many exper-imental studies deriving primary photodissociation quantumyields and product yields following the photolysis of pyruvicacid. The studies which were carried out at atmosphericallyrelevant wavelengths (i.e. within the ∼ 300–400 nm absorp-tion band) are listed in Table S1 in the Supplement. The ex-periments were carried out at different pressures of variousbath gases and at different wavelengths and at different con-centrations of pyruvic acid, all of which appear to play a rolein determining the products formed. Table S1 shows that CO2is formed at a yield of close to 100 %, whereas the yield ofCH3CHO is highly variable. CH2=CHOH has been detectedboth at a few torr of helium (Samanta et al., 2021) and at 1 barof air (Calvert et al., 2011). Other end products observed dur-ing the photolysis of pyruvic acid in 1 bar of air include aceticacid (Calvert et al., 2011; Reed Harris et al., 2016, 2017b)and PAN (Grosjean, 1983; Berges and Warneck, 1992) whenNO2 was present, which together provide evidence for theformation of the acetyl peroxy radical (CH3C(O)O2) andthus CH3CO, for example, in Reactions (R3) and (R4) whensunlight or solar-simulating light sources are used. Sec-ondary products (resulting, for example, from the further re-actions of CH3CHO) such as HCHO and CH3OH have alsobeen observed at pressures close to 1 bar (Grosjean, 1983;Calvert et al., 2011; Reed Harris et al., 2016; Reed Harriset al., 2017a). While the Norrish type 1 process (C–C bondfission) forming CH3CO and HOCO appears to be unimpor-tant at 351 nm (Samanta et al., 2021), it may be favouredat wavelengths < 340 nm (Chang et al., 2014). This is illus-trated in Fig. 1 where we present the wavelength-resolvedphotolysis rate constants across the UV absorption spectrumof pyruvic acid (assuming an overall photolysis quantumyield of 1, and absorption cross-sections recommended byIUPAC). The wavelength-resolved actinic flux was calcu-lated for the IBAIRN measurement site on the 13 Septem-ber 2016 using the Tropospheric Ultraviolet and VisibleRadiation model (TUV; https://www.acom.ucar.edu/Models/TUV/Interactive_TUV/, last access: 1 September 2020). In-tegration of the J values at wavelengths< 340 nm indicatesthat (at local noon) ≈ 25 % of pyruvic acid dissociation oc-curs at these shorter wavelengths.

2 Methods

The goal of this study is to evaluate the impact of pyruvicacid on acetaldehyde and radical formation rates in the bo-real forest using a data-constrained, chemical box model. Forthis purpose we make use of experimental data from a fieldstudy, which was performed in the Finnish boreal forest at

https://doi.org/10.5194/acp-21-14333-2021 Atmos. Chem. Phys., 21, 14333–14349, 2021

14336 P. G. Eger et al.: Impact of pyruvic acid photolysis on CH3CHO and RO2

the Station for Measuring Ecosystem-Atmosphere RelationsII (SMEAR II) in Hyytiälä (61.846◦ N, 24.295◦ E; 180 mabove sea level; see Hari and Kulmala, 2005), an area thatis characterized by large emission rates of biogenics (mainlymonoterpenes) and low-NOx concentrations (Rinne et al.,2000; Williams et al., 2011; Aalto et al., 2015; Fischer et al.,2021).

The variability in the reported photodissociation quantumyield and product distributions (see discussion above) sug-gests that pyruvic acid photodissociation is not yet fully un-derstood, In addition, the fate of the potentially dominant car-bene product (Samanta et al., 2021) is unknown. Therefore,in order to better constrain the fate of CH3COH in the atmo-sphere, quantum chemical calculations were undertaken tocharacterize its likely atmospheric reactions, for which ex-perimental data do not exist.

2.1 The IBAIRN campaign

The IBAIRN campaign took place in September 2016, dur-ing the summer–autumn transition, and was characterizedby frequent temperature inversions near ground level duringnight-time (Liebmann et al., 2018a), which led to the accu-mulation of nocturnally emitted trace gases from vegetation.A detailed description of the campaign and the instrumentsdeployed can be found elsewhere (Liebmann et al., 2018a;Eger et al., 2020). A summary (with details of detection limit,etc.) is provided in Table S2. Briefly, pyruvic acid was mea-sured by a chemical ionization quadrupole mass spectrome-ter (Eger et al., 2020), the sum of monoterpenes (henceforthreferred to as MT) was measured by a PTR-ToF-MS, andsingle MTs were monitored by a GC-AED (Liebmann et al.,2018a). Despite some discrepancies related to instrument lo-cation and inhomogeneity in terpene emissions within theforest, both instruments were in reasonably good agreementthroughout the campaign. Since a high temporal resolutionis preferable for our simulation, we have used the PTR-ToF-MS dataset. NO and NO2 were measured by a chemilumines-cence detector and a cavity ring-down spectrometer (Soban-ski et al., 2016; Liebmann et al., 2018b), ozone was measuredby optical absorption, and CO was measured by quantum-cascade-laser absorption spectroscopy (Eger et al., 2020).Formic and acetic acid as well as methyl ethyl ketone (MEK)and methyl vinyl ketone (MVK) were taken from the contin-uous PTR-MS measurements at the site at heights between42 and 336 m. Photolysis rate coefficients were derived usingactinic flux measurements from a spectral radiometer (MET-CON GmbH) (METCON GmbH) and evaluated cross sec-tions and quantum yields (Burkholder et al., 2015). Mixinglayer (MXL) heights were derived by combining in situ mea-surements made by a scanning Doppler lidar (Hellén et al.,2018) with results from the ECMWF ERA-Interim reanalysis(Dee et al., 2011), with a spatial resolution of∼ 80 km. Sincethe lidar was unable to resolve MXL heights< 60 m (as regu-larly experienced during nocturnal inversions), all values be-

low this threshold have been set to 60 m, representing an up-per limit.

2.2 Theoretical analysis of the fate of singletmethylhydroxy carbene, CH3COH

We investigated the reactions of CH3COH theoreticallyunder atmospheric conditions, examining its unimolecularreactions and bimolecular reactions with O2, H2O, andHC(O)OH, where the latter is representative of carboxylicacids. The reaction with pyruvic acid itself is also briefly ex-plored. The rovibrational characteristics and energetics of allcritical points on the potential energy surface were charac-terized at the CCSD(T)/aug-cc-pVTZ//M06-2X-D3/aug-cc-pVTZ level of theory with a wavenumber scaling factor of0.971 (Zhao and Truhlar, 2008; Dunning, 1989; Purvis andBartlett, 1982; Grimme et al., 2011; Database of FrequencyScale Factors for Electronic Model Chemistries (Version 4);Alecu et al., 2010). This method compares favourably withthe more rigorous focal point analysis of Schreiner et al.(2011), with energy differences in the singlet state unimolec-ular chemistry of less than 0.7 kcalmol−1, indicating that themethod is reliable for kinetic predictions under atmospherictemperatures. Where necessary, broken-symmetry SCF (self-consistent field) calculations were used to describe singletbiradicals (Noodleman, 1981), and IRC (intrinsic reactioncoordinate) calculations were used to verify the pathways.For reactants, products, and transition states, we exhaustivelycharacterized all conformers; for complexes we only char-acterized those directly connecting to a transition state. Allquantum chemical calculations were performed using theGaussian-16 program suite (Frisch et al., 2016).

The quantum chemical data were then used to calculatehigh-pressure rate coefficients for reactions over a saddlepoint using multi-conformer transition state theory (MC-TST) calculations (Vereecken and Peeters, 2003), under arigid rotor harmonic oscillator approximation. Tunnellingcorrections are performed assuming an asymmetric Eckartbarrier (Eckart, 1930; Johnston and Heicklen, 1962). Mostreactions have high energy barriers, and the presence of pre-and post-reaction complexes has a negligible influence on thereaction rate. For barrierless reactions, typically complexa-tion reactions, we assume the reaction rate is close to thecollision limit unless indicated otherwise.

2.3 Box model

We have used the CAABA/MECCA atmospheric chemistrybox model to numerically simulate the impact of pyruvic acidphotolysis on the formation of radicals and CH3CHO overthe diel cycle during the IBAIRN campaign. Our study isbased on model version 4.4.2, with updated reactions relatedto pyruvic acid in which two different scenarios were inves-tigated (see Sect. 3.3) in order to examine the sensitivity of

Atmos. Chem. Phys., 21, 14333–14349, 2021 https://doi.org/10.5194/acp-21-14333-2021

P. G. Eger et al.: Impact of pyruvic acid photolysis on CH3CHO and RO2 14337

the model output to, for example, photolysis quantum yieldsand products.

The chemical mechanism used in this study con-tains > 600 gas-phase species and ∼ 2000 gas-phase re-actions and photolysis steps. In addition to the basicozone, HOx and NOx chemistry, the mechanism containsthe detailed “Mainz Organic Mechanism” (MOM) for non-methane hydrocarbons (NMHC), isoprene, terpenes, and aro-matics. MOM is derived from a reduced version of theMaster Chemical Mechanism (MCM). Full details aboutCAABA/MECCA and MOM are available in Sander et al.(2019). Photolysis reactions are calculated for a latitude of62◦ N. A complete reaction scheme and source of rate coeffi-cients can be found in the data archive (see Data availabilitysection).

Several parameters (temperature, pressure, relative humid-ity) and trace-gas concentrations (pyruvic acid, O3, NO,NO2, PAN, CO, MTs, formic and acetic acid, methyl ethylketone (MEK), and methyl vinyl ketone, MVK) as well asthe photolysis rate constants of various trace gases were con-strained to values measured during the IBAIRN campaign.

Based on the GC-AED measurements, the MTs were splitinto α-pinene (49 %), β-pinene (13 %),1-carene (27 %), andcamphene (8 %). Limonene is not included in the standardchemical mechanism of CAABA/MECCA, but as its contri-bution to the MTs during IBAIRN was only 3 % it was treatedas 1-carene (increasing the 1-carene contribution to 30 %).

The atmospheric methane mixing ratio was set to a con-stant value of 1.8 ppmv. Non-methane alkanes, the degrada-tion of which represents ∼ 30 %–45 % of the acetaldehydesource globally (Millet et al., 2010), were constrained to1000 pptv of ethane, 250 pptv of propane, and 150 pptv of n-butane, as found in similar environments in Finland (Hakolaet al., 2006; Hellén et al., 2015). The mixing ratio of PAN,which is generally the most abundant of the peroxy acetyl ni-trates (PNs), was calculated from a measurement of the sumof peroxy nitrates, whereby [PAN]= 0.9×6[PNs] (estima-tion based on observations by, for example, Shepson et al.,1992b; Roberts et al., 2004; Roiger et al., 2011). The model-generated, averaged OH concentration through the diel cy-cle was in good agreement (within ∼ 20 %) with that cal-culated from the correlation of ground-level OH measure-ments with UVB radiation intensity at the Hyytiälä site (with[OH]= 5.62× 105 [UVB]0.62 moleculecm−3 when UVB isin units of Wm−2; Petäjä et al., 2009; Hellén et al., 2018)but showed more variability resulting from changes in NOmixing ratios and the conversion of HO2 to OH.

3 Results and discussion

In the following, we analyse in situ measurements of pyruvicacid to derive emission rates, present the results of the the-oretical calculations of the fate of CH3OH, and discuss thebox-model output for the IBAIRN campaign with a focus on

pyruvic acid emission rates and its impact on acetaldehydeand radical chemistry in the boundary layer of the boreal for-est.

3.1 Pyruvic acid emission rate relative tomonoterpenes during IBAIRN

In order to derive the pyruvic acid emission rate (Epyr) duringIBAIRN, we assume that only photolysis and dry depositioncontribute significantly to its overall loss rate and that pyru-vic acid is in steady state. The latter assumption is reasonableas its mean lifetime was (2± 0.5) h. Due to a homogeneousfetch at the measurement site, we can neglect transport pro-cesses, and Epyr is defined by Eq. (1), where [pyr]ss is themeasured mixing ratio, Jpyr is the photolysis rate constant ofpyruvic acid, kdep is the first-order loss rate constant for itsdry deposition, and hMXL is the well-mixed boundary layerheight.

Epyr = [pyr]ss(Jpyr+ kdep)hMXL (1)

Epyr is effectively an emission rate normalized to the MXLheight (hMXL). As the photolysis is a substantial fraction ofthe overall losses of CH3C(O)C(O)OH, the choice of quan-tum yield ϕ directly impacts the calculated emission rate.

The deposition rate of pyruvic acid was calculated fromkdep= vdeph

−1MXL during the day and kdep= 2vdeph

−1MXL dur-

ing the night (Shepson et al., 1992a), with the transition fol-lowing the diel variation in the mixing layer height hMXL(see Fig. S1 in the Supplement). Further, as pyruvic acid andH2O2 have similar solubilities, we assumed that their de-position velocities are equal, so that vdep= 8.4 cms−1 dur-ing the day and vdep= 0.8 cms−1 during the night, as de-rived by Crowley et al. (2018) for the same site. This re-sulted in a minimum dry-deposition loss rate constant ofkdep= 0.9× 10−4 s−1 during the day and a maximum ofkdep= 1.8× 10−4 s−1 during the night.

The same calculation is performed for the MTs (EMT) overthe same period (and thus for the same MXL height). We notethat hMXL controls not only the value of kdep, but also directlyaffects the mixing ratios of both MTs and pyruvic acid for agiven emission rate. The relative emission rate (Epyr/EMT)can be calculated from Eq. (2), where terms in square brack-ets are concentrations.

Epyr

EMT=

[pyr]ss(Jpyr+ kdep)

[MT]ss(kOH [OH]+ kNO3 [NO3]+ kO3 [O3]

) (2)

In the denominator, kOH, kNO3 , and kO3 are rate coeffi-cients for reaction of MTs with OH, NO3, and O3, respec-tively. As we do not have GC data at high time resolution,an effective rate coefficient for loss of the monoterpenes wasderived from the mean MT composition as measured by GC-AED (49 % α-pinene, 13 % β-pinene, 27 % carene (sum of2-carene and 3-carene), 3 % d-limonene, and 8 % camphene)and the corresponding rate coefficients (Perring et al., 2013;

https://doi.org/10.5194/acp-21-14333-2021 Atmos. Chem. Phys., 21, 14333–14349, 2021

14338 P. G. Eger et al.: Impact of pyruvic acid photolysis on CH3CHO and RO2

Figure 2. Diel variation of the (MXL height-corrected) emissionrates of pyruvic acid (Epyr, scenario B) and monoterpenes (EMT)along with Jpyr (yellow shaded), T , and hMXL for the IBAIRNcampaign.

Gaona-Colman et al., 2017; IUPAC Task Group on Atmo-spheric Chemical Kinetic Data Evaluation, 2021). This willintroduce significant uncertainty (factor of ∼ 2) into the cal-culation of the MT emission rates. Further uncertainty arisesfrom the measurement of pyruvic acid, 6MT, OH, O3, NO3,hMXL, and Jpyr. In particular, the results are very sensitive tothe deposition velocity (vdep) of pyruvic acid, which is an es-timate based on the deposition velocity of H2O2, which itselfhas an uncertainty of ∼ 90 % (Fischer et al., 2019). Further,our calculations are based on the assumption that the sourcesof pyruvic acid and MT are evenly distributed, and measure-ments made at∼ 8.5 m above the ground are representative ofthe entire boundary layer (i.e. that the boundary layer is well-mixed, including the very shallow boundary layer at night).A gradient in pyruvic acid mixing ratios at night cannot beruled out, which would impact on our results. We estimatethat the emission ratio (Epyr/EMT) in Eq. (2) is associatedwith an overall uncertainty of a factor of ∼ 3 notwithstand-ing the use of different quantum yields (and thus J values)for pyruvic acid photolysis.

A time series of pyruvic acid and MT mixing ratios alongwith the MXL height (hMXL) derived from a lidar mea-surement and from the ERA-Interim reanalysis is shown inFig. S2. Whereas both MXL height datasets agree very wellduring the night when the MXL is shallow (usually< 100 m),the lidar data are on average a factor of ∼ 2 lower duringthe day and characterized by a much higher variability. Forthe derivation of the diel profile of hMXL (Fig. S1), we tookan average of both datasets. The diel variation displayed inFig. S2, with the highest MT mixing ratios at night, is char-acteristic of this boreal forest site and has been observed inearlier studies (Hellén et al., 2018).

In the following, we focus on the mean, diel profiles ofEpyr, EMT, Jpyr, T , and hMXL for the IBAIRN campaign,

which are presented in Fig. 2. A plot showing the variabilityof the MT and pyruvic acid mixing ratios over the same pe-riod was previously reported (see Fig. 3 of Eger et al., 2020).

During September, the emission rate of pyruvic acid (Epyr)reaches its maximum a few hours after solar noon whenthe temperature peaks, similar to EMT. However, the ampli-tude of the day-to-night difference in Epyr is a factor of ∼ 3smaller than observed forEMT. This could indicate that pyru-vic acid emissions are less temperature-dependent than MTemissions (see below) and that other environmental factorsmight additionally play a role at this time of year.

The emission rates of the MTs derived as described aboveshow a large day–night variation with values a factor of∼ 20 larger around noontime compared to midnight. Thisis significantly larger than the expected variation (factorof 2–3) based on the average noon-to-midnight tempera-ture difference of 10 K and the parameterization of Guen-ther et al. (1993), whereby EMT ∝ exp(β(T − 297K)) withβ = 0.1 K−1 (which is in line in with the empirical value ofβ = 0.12 K−1 that was derived for this site in September byHellén et al., 2018). One potential reason for this discrep-ancy may be emissions in autumn from fresh leaf litter thatsignificantly contribute to the observed mixing ratios (Hellénet al., 2018) although the assumption of evenly distributedsources and a well-mixed boundary layer is not necessarilyvalid during the night, especially during strong temperatureinversions.

Figure S3 shows that the daytime emission of pyruvic acidrelative to MT (Epyr/EMT) varies by a factor of∼ 2, depend-ing on the chosen scenario, whereas the nighttime emissionratio is only dependent on the deposition velocity of pyruvicacid. For further analysis we adopt a quantum yield of 0.2,as presently preferred by IUPAC. On average, (Epyr/EMT)∼ 0.6, with a minimum value of ∼ 0.3 in the evening and amaximum value of ∼ 1 in the early morning, indicating ele-vated pyruvic acid emissions relative to MT at night. To de-rive a T -dependent expression from the diurnal profile of theemission factor, we fit an exponential function to the plot oftemperature versus Epyr/EMT (Fig. S4), yielding

Epyr =

[0.28+ 3.17 · exp

(273− T

4.24

)]·EMT. (3)

We note that (like the values of Epyr) the temperature de-pendence derived is strongly influenced by the diel variationof the MXL height and thus carries significant uncertaintyand may not be transferable to other locations or even timesof the year.

As our measurements of pyruvic acid are the first to havebeen made in the boreal forest, we cannot compare our rela-tive emission ratio (Epyr/EMT) with previous measurementsin a similar environment. Instead, where possible, we derivethe emission ratio from measurements of MTs, isoprene, andpyruvic acid in warmer climates.

Jardine et al. (2010b) performed measurements in an en-closed (glass dome) tropical forest biome at Biosphere 2 in

Atmos. Chem. Phys., 21, 14333–14349, 2021 https://doi.org/10.5194/acp-21-14333-2021

P. G. Eger et al.: Impact of pyruvic acid photolysis on CH3CHO and RO2 14339

Table 1. Emission rate of pyruvic acid (Epyr) relative to isoprene (Eiso) and MT (EMT).

Reference Location Plant species (Epyr/Eiso) (Epyr/EMT)

This study Hyytiälä, Finland Boreal forest ∼ 20 0.62Talbot et al. (1990) Manaus, Brazil Tropical forest 0.003 –Jardine et al. (2010b) Biosphere 2, Arizona, US Tropical biome 0.17 4Jardine et al. (2010b) Biosphere 2, Arizona, US Mango tree 0.05 1.7Jardine et al. (2010a) Biosphere 2, Arizona, US Creosote bush 0.05 0.07

Arizona, US, where they found maximum mixing ratios of120 ppbv isoprene, 6 ppbv MTs, and 15 ppbv pyruvic acid.As the glass dome absorbed actinic wavelengths and pre-vented active photochemistry, the chemical loss processesfor pyruvic acid, isoprene, and MT (including photolysisand reactions with OH, O3, and NO3) are negligible. Ini-tially disregarding the deposition of isoprene and MT, wederive lower limits of (Epyr/Eiso) ∼ 0.17 and (Epyr/EMT)∼ 4 (see Table 1). However, due to the presence of largeconcentrations of isoprene-consuming microbes in the soilof Biosphere 2, the isoprene loss rate via deposition maybe enhanced, which will decrease the effective emission ra-tio (Epyr/Eiso). In addition, branch enclosure studies wereperformed on a Mangifera indica (mango) tree within Bio-sphere 2, yielding mean fluxes (in nmolm−2 s−1) of 3.2 forisoprene, 0.09 for MT, and 0.15 for pyruvic acid. Pyruvicacid emissions peaked during the day when temperature andphotosynthetically active radiation (PAR) were highest andcorrelated very well with isoprene emissions and (to a certainextent) with MT emissions. Assuming that a mango tree isrepresentative of the tropical vegetation, we derive an emis-sion ratio of (Epyr/Eiso ∼ 0.05 and (Epyr/EMT) ∼ 1.7 (seeTable 1), which is consistent with our estimations for theIBAIRN campaign. However, given that Talbot et al. (1990)observed great variability in pyruvic acid emission fluxesamong five different tree species during measurements inthe tropical Ducke Forest Reserve close to Manaus, Brazil,this agreement may, to some extent, be coincidental. Talbotet al. (1990) also reported a mean emission flux (derived fromenclosure experiments) relative to isoprene of (Epyr/Eiso)∼ 0.003, which is about 1 order of magnitude smaller thanin the study of Jardine et al. (2010b). In a further branchenclosure study by Jardine et al. (2010a) emissions from acreosote bush (Larrea divaricata), which is typically foundin US drylands, were investigated. Average noontime branchemission rates (in µgCgdw−1 h−1) of 7.5, 10.4, and 0.2 forisoprene, MT, and pyruvic acid resulted in relative emissionratios of (Epyr/Eiso) ∼ 0.05 and (Epyr/EMT) ∼ 0.07 for thismixed isoprene–MT-emitting species.

The comparison with the few datasets available in the lit-erature indicates that the variability of the emission factors(Epyr/EMT) and (Epyr/Eiso) among different plant speciesand different environments is large. In addition, a lack ofpyruvic acid measurements over different seasons in the bo-

real forest means that we cannot exclude that the value wederive is biased by emissions (e.g. from ground-level, decay-ing plant litter in September) that are particular to this seasonand environment. The emission rates we derive are thereforerelevant for the autumnal boreal forest but require validationbefore being extended to other regions and seasons with con-fidence.

3.2 Theoretical calculations on the fate of CH3COH

Singlet methylhydroxy carbene, CH3COH, is best character-ized as having an sp2-hybridized central carbon, bearing anin-plane lone pair in an sp2 orbital and an empty p orbital per-pendicular to the CCO plane. The lone pairs of the hydroxy Oatom back-donate into the empty p orbital, such that the mostfavourable geometry has the hydroxy–H atom in the CCOplane. The orientation of the terminal OH group has a largeimpact on the energy, with 3 kcalmol−1 energy differencebetween the syn- and anti-conformers. Due to the interac-tion between the hydroxy O atom and the carbene function-ality, internal rotation of the OH group has a very high bar-rier, 24 kcalmol−1. Concomitantly, syn/anti-interconversionis very slow, with predicted rate coefficients at 300 K of lessthan 10−2 s−1. Under atmospheric conditions, thermalizedsyn- and anti-CH3COH are thus best considered as separatespecies, with possibly distinct chemistry. No information isavailable on the relative yield of these conformers from pyru-vic acid photolysis.

3.2.1 Unimolecular reactions of CH3COH

Both syn- and anti-CH3COH can isomerize to vinyl alco-hol over high barriers ≥ 24 kcalmol−1 (see Fig. 3). Anti-CH3COH has an additional pathway for isomerization to ac-etaldehyde, with a barrier of 23 kcalmol−1. Due to these highbarriers, the thermal rate of isomerization is comparativelyslow, with a 300 K rate coefficient of≤ 4× 10−4 s−1 (see Ta-ble 2). As already discussed by Schreiner et al. (2011), for-mation of CH3CHO from anti-CH3COH is most favourableat low temperatures, owing to a thinner energy barrier andhence faster tunnelling. At temperatures above 260 K, wefind that formation of CH2=CHOH from anti-CH3COH be-comes dominant, with a ∼ 3.5 : 1 ratio of CH2=CHOH toCH3CHO at room temperature.

https://doi.org/10.5194/acp-21-14333-2021 Atmos. Chem. Phys., 21, 14333–14349, 2021

14340 P. G. Eger et al.: Impact of pyruvic acid photolysis on CH3CHO and RO2

Table 2. Theory-predicted high-pressure rate coefficients for reaction of singlet CH3COH.

Reactants Products k(298 K) A n Ea

syn-CH3COH+O2 CH3C(OH)OO q 2.2× 10−20 5.74× 10−22 3.05 4092

anti-CH3COH+O2 CH3C(OH)OO q 6.6× 10−21 1.71× 10−22 2.97 3960

syn-CH3COH+H2O CH3CH(OH)2 1.9× 10−20 1.57× 10−55 13.56 −1049

anti-CH3COH+H2O CH3CH(OH)2 5.7× 10−21 1.09× 10−61 15.61 −1443

syn-CH3COH anti-CH3COH 8.9× 10−3 7.86× 10−20 10.77 6598CH2=CHOH 1.9× 10−4 3.62× 10−91 34.20 −1444

anti-CH3COH syn-CH3COH 2.8× 10−5 6.55× 10−20 10.71 8137CH2=CHOH 9.2× 10−5 2.02× 10−14 40.40 −6660CH3C(=O)H 3.4× 10−4 1.26× 10−81 30.96 −563

CH2=CHOH+HCOOH CH3C(=O)H+HCOOH 2.9× 10−18 1.82× 10−76 19.88 −6192

CH3C(=O)H+HCOOH CH2=CHOH+HCOOH 8.1× 10−27 1.09× 10−78 20.59 −633

Calculations were performed at the CCSD(T)//M06-2X-D3 with MC-TST level of theory. Rate coefficients are given at 298 K (s−1 orcm3 molecule−1 s−1). Temperature-dependent rate coefficients can be calculated using the parameters of a Kooij expressionk(200-450 K)=A · (T /K)n · exp(−Ea/T ) with A given per second (s−1) or cubic centimetres per molecule per second(cm3 molecule−1 s−1) and Ea in kelvin (K).

Figure 3. Zero point energy (ZPE)-corrected potential energysurface for unimolecular reactions of singlet CH3COH at theCCSD(T)//M06-2X-D3 level of theory.

Given the low predicted thermal rate coefficients, it seemsunlikely that the experimentally observed acetaldehyde andvinyl alcohol in pyruvic acid photolysis are formed from iso-merization of thermalized CH3COH. The energy distribu-tion of energized, nascent carbenes would be rather broadas the available energy upon pyruvic acid photodissocia-tion is distributed over all fragments and their relative mo-tion, and the isomerization yield would then be pressure-dependent. The CH3CHO and CH2=CH2OH isomers formedwould have enough energy to undergo keto-enol tautomeriza-tion, but given the high barrier exceeding 55 kcalmol−1, it is

Figure 4. ZPE-corrected potential energy surface for reaction ofsinglet CH3COH with O2 at the CCSD(T)//M06-2X-D3 level oftheory.

more probable they will instead be stabilized by collisionalenergy loss.

3.2.2 Reaction of CH3COH with O2

Under atmospheric conditions, the reaction with O2 is poten-tially an important loss process for CH3COH (Reed Harriset al., 2016, 2017a; Eger et al., 2020). The potential energysurface is shown in Fig. 4. Contrary to radicals, which re-act with O2 by (near-)barrierless radical recombination, thesinglet CH3COH carbene does not have an unpaired electronand the reaction proceeds mostly by association of its out-of-plane empty p orbital with a lone electron pair in O2, re-

Atmos. Chem. Phys., 21, 14333–14349, 2021 https://doi.org/10.5194/acp-21-14333-2021

P. G. Eger et al.: Impact of pyruvic acid photolysis on CH3CHO and RO2 14341

quiring orbital rearrangement to a triplet C qOO q moiety witha sp3-hybridized central carbon. This unfavourable processhas high barriers,> 9 kcalmol−1, and concomitantly low ratecoefficients, k(298 K) ∼ 10−20 cm3 molecule−1 s−1 (see Ta-ble 2). The rate coefficient is however highly uncertain, ow-ing to an uncertainty (∼ 1 to 2 kcalmol−1) on the barrierheight.

The decomposition of the CH3C q(OH)OO q triplet dirad-

ical intermediate, forming CH3C q=O+HO2, is reminiscentof the chemistry of α-OH alkyl radicals with unpaired elec-trons and should occur rapidly owing to the sufficientlyhigh energy content of the peroxyl-alkyl diradical (Her-mans et al., 2005, 2004; Dillon et al., 2012; Olivella et al.,2001; Dibble, 2002). Note that this chemistry is very distinctfrom that of the singlet CH3C(OH)OO Criegee intermediate.The acyl radical product is expected to recombine rapidlywith a second O2 molecule, forming acylperoxy radicals,CH3C(=O)OO q. Alternatively, the triplet CH3C q

(OH)OO qintermediate can react with a second O2 molecule by a bar-rierless recombination reaction (Fig. 4), forming the diper-oxy singlet diradical CH3C(OH)(OO q

)OO q, which in turncan eliminate HO2, similar to other α-OH peroxy radicals,forming the acylperoxy radicals directly. This second O2 ad-dition is sufficiently exothermic to allow for formation of per-acetic acid with a singlet O2 molecule, but this process has arather large barrier of∼ 24 kcalmol−1 and is expected to be aminor contributor, leaving CH3C(O)O2+HO2 as the likelydominant products of the overall reaction of CH3COH withoxygen molecules.

3.2.3 Reactions of CH3COH with carboxylic acids

Samanta et al. (2021) observed loss of CH3COH via reac-tion with pyruvic acid, which may indicate that its fate in theatmosphere may also be (partially) controlled by similar re-actions. To theoretically investigate the reaction of CH3COHwith carboxylic acids, we used formic acid in the calcula-tions. Not only is formic acid an abundant organic acid in theatmospheric boundary layer, but also its reactivity is relatedto the properties of the –C(=O)OH moiety, and the results aretransferable to other oxoacids, including pyruvic acid, whichwas present in high concentrations in most laboratory inves-tigations.

As shown in Fig. 5, CH3COH forms strong complexeswith HC(O)OH, with 11 kcalmol−1 stability. From this com-plex, an addition process occurs that is best described as thetransfer of the acidic H+ atom to the carbene lone pair onthe CH3COH central carbon, with simultaneous associationof one of the negatively charged lone electron pair of the car-bonyl oxygen to the carbene vacant p orbital, forming a 1-hydroxyethyl ester (CH3CH(OH)OC(O)H). Due to the con-certed association of the two carbene orbitals with suitablepartners in the carboxylic moiety, this process has a very lowbarrier (≤ 1 kcalmol−1). This mechanism is feasible due tothe size of the –C(O)OH group and the possibility of shift-

Figure 5. ZPE-corrected potential energy surface for reactionsof singlet CH3COH with H2O (left) and HCOOH (right) at theCCSD(T)//M06-2X-D3 level of theory.

ing the double bond to the other oxygen atom upon H-atomloss. For the anti-CH3COH carbene, we also found that anin-plane approach of the carboxylic acid towards the COHmoiety in methylhydroxy carbene can simultaneously trans-fer the acidic H atom to the carbene carbon while the car-bene hydroxy H atom is transferred to the carbonyl oxygen inthe acid, reforming the HC(O)OH co-reactant. This catalysisreaction converts anti-CH3COH to acetaldehyde, CH3CHO,without an energy barrier. Both adduct formation and thecatalysis reaction should proceed with rate coefficients nearthe collision limit.

Carboxylic acids can also catalyse keto-enol tautomeriza-tion, possibly helping the isomerization between CH3CHOand CH2=CH2OH by reducing the effective barrier by over50 kcalmol−1 though the thermal reaction remains slow (seeTable 2). The only reaction of CH3COH that has been in-vestigated experimentally to date is that with pyruvic acid(Samanta et al., 2021), which we also examine theoreti-cally in the Supplement. Note that the large rate coefficientfor CH3COH with organic acids calculated here would im-ply that reaction of thermalized CH3COH with pyruvic acidwould overwhelm any other bimolecular CH3COH reactionin their work and most of the experiments listed in Table S1.

3.2.4 Reactions of CH3COH with H2O

Based on the reactivity of small carbenes towards closed-shell molecules, Samanta et al. (2021) suggested that reac-tion with H2O might be an important loss process of theCH3COH carbene intermediate. We have characterized theinsertion reaction of CH3COH in the H2O molecule andfound very high barriers,≥ 11 kcalmol−1, with very low ratecoefficients ∼ 10−20 cm3 molecule−1 s−1 (see Fig. 5 and Ta-ble 2). The reaction is significantly slower than with car-boxylic acid as the smaller H2O molecule is unable to simul-taneously reach both carbene orbitals in a favourable geome-try. The reaction of H2O with CH3COH is best described as ashift of an H+ atom to the carbene lone pair orbital, followed

https://doi.org/10.5194/acp-21-14333-2021 Atmos. Chem. Phys., 21, 14333–14349, 2021

14342 P. G. Eger et al.: Impact of pyruvic acid photolysis on CH3CHO and RO2

by migration of the water HO− moiety to the vacant carbeneorbital to form a bond with a lone electron pair. The reac-tion is further hindered by the back-donation of the CH3COHoxygen atom into the vacant carbene orbital, partially fillingthe vacant carbene orbital and reducing the reactivity of thecarbene functionality. We therefore propose that CH3COHwill be significantly less reactive towards closed shell speciesthan the 2CH and 1CH2 carbenes which are known to exhibitvery fast insertion and cyclo-addition reactions (Vereeckenet al., 1998; Goulay et al., 2009; Douglas et al., 2019; Jasperet al., 2007; Gannon et al., 2010).

3.2.5 Summary of theoretical calculations: the fate ofthermalized CH3COH in 1 bar of air

The theoretical analysis of the fate of CH3COH carbene in-termediates formed in PA photolysis indicates that the ac-etaldehyde formation observed in many experiments couldbe the result of a fast catalysis reaction of CH3COH withpyruvic acid, which under typical experimental conditionsdominates over competing reactions, such as with O2, byseveral orders of magnitude. This conclusion is consistentwith the experimental observations of Reed Harris et al.(2017a), who report a reduction in the acetaldehyde yieldwhen low pyruvic acid concentrations are used and an in-crease in the formation of acetic acid (which can be formedin the reaction of CH3C(O)O2 radicals with HO2). In theatmospheric boundary layer atmosphere, where the con-centrations of organic acids may lie between 1010 and1011 moleculecm−3 (Millet et al., 2015) and that of O2 isclose to 5× 1018 moleculecm−3, the reactions of CH3COHwith organic acids and O2 are competitive, whereas reactionof CH3COH with water is minor. Table 2 lists the predictedrate coefficients for these reactions.

3.3 Box-model results: contribution of pyruvic acid toacetaldehyde and radical formation

To account for the large variability in photodissociationquantum yields and product yields reported in the literature(see above), we modelled two scenarios, A and B:

– Scenario A. In this scenario we used pyruvic acid crosssections, quantum yields, and product yields accord-ing to the IUPAC recommendations with a photodis-sociation quantum yield (ϕ) of 0.2 at 1 bar pressureand branching ratios of 0.6, 0.05, and 0.35 for Reac-tions (R1), (R2), and (R3) as listed in Sect. 1.1.

– Scenario B. Here we use the same absorption cross-sections as scenario A but build on the recent ob-servations of (Samanta et al., 2021) and the theoreti-cal work presented in Sect. 3.2, which considers theformation and fate of an excited CH3COH molecule(+CO2). In scenario B, we consider the effects of us-ing photodissociation quantum yields of 0.2, 0.5, and

1 (scenarios B0.2, B0.5, and B1, respectively). Photol-ysis at wavelengths < 340 nm was considered to gen-erate CH3CO+HOCO, whereas photolysis at wave-lengths > 340 nm was assumed to form CO2+ energyrich CH3COH#, which undergoes the reactions outlinedin Sect. 1. Assuming a quantum yield that is indepen-dent of wavelength results in 25 % of pyruvic acid pho-tolysis at noon taking place at wavelengths < 340 nmand 75 % at wavelengths > 340 nm. In the model, weassume that this ratio does not change (i.e. we ne-glect wavelength-dependent variations in the relativeactinic flux through the diel cycle). The values of 25 %and 75 % listed above roughly correspond to the rela-tive importance of peroxy radical formation (via Reac-tions R3, R4, and R5) at the shorter wavelengths com-pared to CH3COH+CO2 formation (Reaction R6) atthe longer wavelengths. Some experimental data indi-cate that addition of O2 can reduce the CH3CHO yieldin favour of formation of, for example, acetic acid. Forthis reason we use a rate coefficient for reaction ofCH3COH with O2 that is competitive with the reactionbetween CH3COH and organic acids. This is a factor of∼ 10 larger than the value obtained theoretically, but weconsider this value still within the uncertainty (∼ 1 to2 kcalmol−1 on the barrier height) of our current theo-retical results as the peculiar wavefunction of CH3COHmay require even higher levels of theory to be describedaccurately.

In the box model, in addition to reaction with O2, the ther-malized carbene also reacts with formic and acetic acids toform acetaldehyde:

CH3COH+HCOOH→ CH3CHO+HCOOH (R12)CH3COH+CH3C(O)OH→ CH3CHO+CH3C(O)OH (R13)

with a rate coefficient of 5× 10−11 cm3 molecule−1 s−1.We assumed that (at 1 bar) 70 % of CH3COH# was

quenched to CH3COH, 20 % isomerized to CH3CHO, and10 % isomerized to CH2=CHOH in order to reproduce theCH3CHO-to-CH2=CHOH ratio reported by Samanta et al.(2021). A summary reaction scheme for the photodissoci-ation of pyruvic acid and the fate of the initial products isgiven in the SI.

3.3.1 CH3CHO formation

The modelled formation of CH3CHO from pyruvic acid pho-tolysis through the diel cycle when considering scenario Ais displayed as a stacked plot of contributing reactions inFig. 6. Immediately apparent from this figure is the domi-nance of pyruvic acid photolysis compared to all other pro-cesses. Under scenario A, even with the low quantum yield(ϕ= 0.2) recommended by IUPAC, pyruvic acid photolysiscontributes > 80 % to the overall CH3CHO production term,with a maximum of ∼ 15 % (at noon) arising from reactions

Atmos. Chem. Phys., 21, 14333–14349, 2021 https://doi.org/10.5194/acp-21-14333-2021

P. G. Eger et al.: Impact of pyruvic acid photolysis on CH3CHO and RO2 14343

Figure 6. Modelled rates of CH3CHO formation (in pptv s−1) through the diel cycle from photolysis of pyruvic acid (blue, orange, andgreen) and other reactions during IBAIRN. (a) Scenario A (IUPAC). (b) Scenario B1. (c) Scenario B0.5. (d) Scenario B0.2. In the legend, thefirst term is the equation tag used by CAABA/MECCA for the reaction. LC4H9O2 is the peroxy radical formed in the reaction of OH withbutane. A full listing of the reactions can be downloaded (see Data availability section).

of the ethylperoxy radical, formed in the reaction of OH withethane (7.5 %) and butane (3.75 %) and in the photolysis ofCH3C(O)C2H5 (3.75 %).

Under scenario B, pyruvic acid photolysis still dominatesthe formation of CH3CHO, with a noontime contribution of91 %, 86 %, and 71 % when quantum yields of 1, 0.5, and0.2 are considered. Of the pyruvic acid contribution, 45 % ofthe CH3CHO arises via isomerization of the initially formed,energized carbene (blue), while the remaining 55 % resultsfrom reactions of the thermalized carbene with formic (or-ange) and acetic (green) acids, the concentrations of whichwere constrained by observations. The modelled, noontimemixing ratio of CH3CHO varies from 400 pptv (scenario B1)to 160 pptv (scenario B0.2) when pyruvic acid photolysis isincluded and is reduced to ∼ 100 pptv when the quantumyield is set to zero. Unfortunately, reliable measurementsof the CH3CHO mixing ratios with which to compare themodel simulations were not available for the IBAIRN cam-paign as the in situ PTR-MS data set (m/z45) varied from−400 to+ 400 pptv over the diel cycle. The modelled, max-imum mixing ratio of CH3CHO increases from ∼ 100 pptvwhen pyruvic acid photolysis is neglected to > 400 pptv un-der scenario B1 (see Fig. S6).

3.3.2 CH3C(O)O2 formation

The CH3C(O)O2 radical is formed in a termolecular reactionbetween the CH3CO radical and O2. Figure 7 displays themain photochemical reactions that lead to the formation of

CH3CO in our model. The spikes in the simulated productionrates are connected to spikes in the diel average NO mixingratio at the site. In analysing the data, we therefore considernot only the contributions at noon (when NO mixing ratioswere large), but also at 10:30 when NO mixing ratios werecomparably low.

Under scenario A, where ϕ= 0.2 and the yield of theCH3CO radical is 0.35, the contribution of pyruvic acid pho-tolysis to the overall production rate at 12:00 and 10:30is about 23 % and 16 %, respectively, roughly equally di-vided into a direct contribution (J43018) and an indirect con-tribution (G42008a) arising via enhanced CH3CHO levels.The main contributors to the formation of CH3CO are re-actions initiated by the degradation of isoprene and MTs(in the legend of Fig. 7: BIACETO2, C511O2, C716O2,CO23C4CHO, CO235C6CHO), which involve reactions ofperoxy radicals with NO.

Under scenario B1, the photolysis of pyruvic acid becomessignificantly more important, contributing a total of 63 % ofthe total production rate for CH3CO at 10:30 and 42 % at12:00. When considering scenarios B0.5 and B0.2 the contri-butions of pyruvic acid photolysis are reduced to 46 % (29 %)and 29 % (17 %), respectively, where the numbers in paren-theses are for the “high-NOx” situation. Generally, the reac-tion of the thermalized carbene with O2 (G42099), the di-rect photolysis at wavelengths < 340 nm (J43018), and theindirect enhancement in CH3CO formation via the enhancedlevels of CH3CHO (G42008a) contribute roughly equally tothe formation of CH3CO resulting from pyruvic acid photol-

https://doi.org/10.5194/acp-21-14333-2021 Atmos. Chem. Phys., 21, 14333–14349, 2021

14344 P. G. Eger et al.: Impact of pyruvic acid photolysis on CH3CHO and RO2

Figure 7. Modelled rates of CH3CO formation (in pptv s−1) through the diel cycle from photolysis of pyruvic acid (blue, orange, and green)and other photochemical processes during IBAIRN. (a) Scenario A (IUPAC). (b) Scenario B1. (c) Scenario B0.5. (c) Scenario B0.2. In thelegend, the first term is the equation tag used by CAABA/MECCA for the reaction. A full listing of the reactions can be downloaded (seeData availability section).

Figure 8. Modelled rates of HO2 formation (in pptv s−1) through the diel cycle from photolysis of pyruvic acid (blue, orange, and green) andother photochemical processes during IBAIRN. (a) Scenario A (IUPAC). (b) Scenario B1. (c) Scenario B0.5. (d) Scenario B0.2. In the legend,the first term is the MCM designation for the reaction. A full listing of the reactions can be downloaded (see Data availability section).

Atmos. Chem. Phys., 21, 14333–14349, 2021 https://doi.org/10.5194/acp-21-14333-2021

P. G. Eger et al.: Impact of pyruvic acid photolysis on CH3CHO and RO2 14345

ysis. The modelled mixing ratio of the CH3C(O)O2 radicalat noon increases by a factor of ∼ 1.5 when comparing sce-nario B1 with the quantum yields for pyruvic acid photodis-sociation set to zero.

3.3.3 HO2 formation

In Fig. 8, we plot the nine most important model pathways toHO2 production through the diel cycle. The dominant mod-elled production terms for HO2 involve HCHO (photolysisHCHO and reaction with OH, G4108, J41001b), the reactionof methoxy radicals (G4118, whereby CH3O is generatedmainly in the reaction of CH3O2 radicals with NO), and thereaction of OH with CO. The direct contribution of pyruvicacid photolysis to HO2 formation (via its photolysis (J43018)and through the reaction of CH3COH with O2, G42099) is∼ 10 % under scenario B1 under low-NOx conditions (i.e. at10:30). Under all other scenarios it is lower, with values (inpercent) of < 1 (scenario A at both 10:30 and 02:00), ∼ 6(scenario B1 at 12:00), ∼ 5 and ∼ 3.5 (scenario B0.5 at 10:30and 12:00, respectively), and ∼ 1.5 and < 1 (scenario B0.2 at10:30 and 12:00, respectively). However, although the directimpact of pyruvic acid photolysis is weak, it has a significantindirect effect via the enhanced formation of CH3C(O)O2radicals (directly via Reactions R3+R4 and R10 and indi-rectly via CH3CHO formation), which, in the presence of O2,reacts with NO to form CH3O2. Enhanced production ratesof CH3O2 result in enhanced production rates of CH3O andHCHO and thus HO2.

CH3C(O)O2+NO(+O2)→ CH3O2+NO2+CO2 (R14)CH3O2+NO→ CH3O+NO2 (R15)CH3O+O2→ HCHO+HO2 (R16)HCHO+hν(+2O2)→ 2HO2+CO (R17)

The model simulations have shown that the photolysis ofpyruvic acid at the levels observed during the IBAIRN cam-paign has a potentially significant effect on both CH3CHOmixing ratios and production rates of HO2 and CH3C(O)O2radicals, the latter being especially enhanced under low-NOxconditions. The enhanced production rates and concentra-tions of CH3C(O)O2 and HO2 also result in significant in-creases in the modelled mixing ratios of several trace gasesthat are formed from these radicals. Comparing scenario B1to the case in which the pyruvic acid photodissociation quan-tum yield (ϕ) is set to zero results in an increase by fac-tors of 2.2, 2.0, and 1.6 for CH3C(O)OOH, CH3OOH, andH2O2, respectively (see Fig. S6). HCHO mixing ratios areenhanced by a factor of 1.2. Vinyl alcohol mixing ratios ofup to 40 pptv were generated in scenario B1. Clearly, the pho-tolysis of pyruvic acid can potentially impact strongly on theconcentrations of, for example, C1 and C2 carbonyl com-pounds and peroxides in the boreal environment.

4 Conclusions

We have combined measurements of pyruvic acid in an au-tumn campaign in the boreal forest (IBAIRN) with theo-retical calculations designed to characterize the fate of themethylhydroxy carbene radical (CH3COH, the major prod-uct of pyruvic acid photodissociation) with a box-modellingstudy. We investigated the impact of pyruvic acid photolysison the rates of production of acetaldehyde (CH3CHO) andthe peroxy radicals CH3C(O)O2 and HO2. The theoreticalstudy revealed unexpected features of CH3COH chemistry,including slow reactions of thermalized carbene with H2Obut an efficient acid-catalysed conversion to CH3CHO in thepresence of organic acids such as HC(O)OH. The reaction ofCH3COH with O2 is slow but will contribute to its fate (andthus the formation of CH3C(O)O2 and HO2) in the loweratmosphere where O2 concentrations are high if the rate con-stant used (elevated by an order of magnitude compared tothe theoretical value) is correct.

In our box model, the photolysis of pyruvic acid was pa-rameterized as presently recommended by IUPAC (wherebythe main products are CH3CHO and CO2) and also using amore detailed mechanism in which the formation and fate ofCH3COH was considered and in which the quantum yieldwas varied. In all scenarios, we find that the photolysis ofpyruvic acid was the dominant source of CH3CHO duringIBAIRN and that its instantaneous contribution to the day-time formation of CH3C(O)O2 varied between 16 % and63 %, dependent on the assumed scenario and also on theNO concentration. Pyruvic acid photodissociation results ina significant increase in the mixing ratios of several carbonylcompounds and peroxides in the boreal environment.

The results of our modelling study are strongly dependenton the chosen quantum yields and photodissociation mech-anism. To reduce the uncertainty in the role of pyruvic acidphotolysis, there is an urgent need for further experimentaland theoretical work on the photochemistry of pyruvic acidand on the fate of methylhydroxy carbene under atmosphericconditions. In addition, further measurements of pyruvic acidmixing ratios and of its deposition velocity in different envi-ronments are required to better constrain its abundance andlifetime and thus the impact of its photolysis. Enclosure stud-ies would be helpful to investigate the dependence of pyruvicacid emission rates on different plant types and environmen-tal conditions.

Data availability. The Max Planck Institute data used for theIBAIRN analysis and the reaction scheme used in the box model arearchived at https://doi.org/10.5281/zenodo.3254828 (Crowley andFischer, 2019).

Supplement. The supplement related to this article is available on-line at: https://doi.org/10.5194/acp-21-14333-2021-supplement.

https://doi.org/10.5194/acp-21-14333-2021 Atmos. Chem. Phys., 21, 14333–14349, 2021

14346 P. G. Eger et al.: Impact of pyruvic acid photolysis on CH3CHO and RO2

Author contributions. PGE was responsible for the pyruvic acidmeasurement during IBAIRN. PGE and JNC, with contributionsfrom JL, wrote the manuscript. LV conducted the theoretical cal-culation on the fate of methylhydroxy carbene, RS and AP did thebox modelling, and NS was responsible for the CRDS measure-ments of NO2 and PANs during IBAIRN. JS was responsible forthe O3 and J value measurements during IBAIRN. HF was respon-sible for the NO and CO measurements during IBAIRN. EK andJW were responsible for the MT measurements during IBAIRN.VV was responsible for the mixing layer height measurements dur-ing IBAIRN. TP was responsible for the SMEAR II observationsand infrastructure. All authors contributed to the paper.

Competing interests. The authors declare that they have no conflictof interest.

Disclaimer. Publisher’s note: Copernicus Publications remainsneutral with regard to jurisdictional claims in published maps andinstitutional affiliations.

Acknowledgements. We thank the technical staff of SMEAR II sta-tion for the excellent support during IBAIRN.

Financial support. This research has been supported by theENVRIplus (grant no. 654182). We are grateful to ENVRIplus forpartial financial support of the IBAIRN campaign.

The article processing charges for this open-accesspublication were covered by the Max Planck Society.

Review statement. This paper was edited by Gabriele Stiller and re-viewed by two anonymous referees.

References

Aalto, J., Porcar-Castell, A., Atherton, J., Kolari, P., Pohja, T.,Hari, P., Nikinmaa, E., Petaja, T., and Back, J.: Onset of pho-tosynthesis in spring speeds up monoterpene synthesis andleads to emission bursts, Plant Cell Environ., 38, 2299–2312,https://doi.org/10.1111/pce.12550, 2015.

Alecu, I. M., Zheng, J. J., Zhao, Y., and Truhlar, D. G.: Com-putational thermochemistry: scale factor databases and scalefactors for vibrational frequencies obtained from electronicmodel chemistries, J. Chem. Theor. Comput., 6, 2872–2887,https://doi.org/10.1021/ct100326h, 2010.

Andreae, M. O., Talbot, R. W., and Li, S. M.: Atmospheric measure-ments of pyruvic and formic acid, J. Geophys. Res.-Atmos., 92,6635–6641, https://doi.org/10.1029/JD092iD06p06635, 1987.

Berges, M. G. M. and Warneck, P.: Product quantumyields for the 350 nm photodecomposition of pyruvic-acid in air, Ber. Bunsen Phys. Chem., 96, 413–416,https://doi.org/10.1002/bbpc.19920960334, 1992.

Burkholder, J. B., Sander, S. P., Abbatt, J., Barker, J. R., Huie, R.E., Kolb, C. E., Kurylo, M. J., Orkin, V. L., Wilmouth, D. M.,and Wine, P. H.: Chemical Kinetics and Photochemical Data forUse in Atmospheric Studies, Evaluation No. 18, JPL Publication15-10, Jet Propulsion Laboratory, Pasadena, available at: http://jpldataeval.jpl.nasa.gov (last access: September 2020), 2015.

Calvert, J. G., Mellouki, A., Pilling, M. J., and Wallington, T. J.:The Mechanisms of Atmospheric Oxidation of the Oxygenates,Oxford Univ. Press, New York, 2011.

Chang, X.-P., Fang, Q., and Cui, G.: Mechanistic photode-carboxylation of pyruvic acid: Excited-state proton transferand three-state intersection, J. Chem. Phys., 141, 154311,https://doi.org/10.1063/1.4898085, 2014.

Crowley, J. N. and Fischer, H.: IBAIRN data (Boreal forest,Hyytiälä Sept. 2016) Max-Planck-Institut, Mainz [data set], Zen-odo, https://doi.org/10.5281/zenodo.3254828, 2019.

Crowley, J. N., Pouvesle, N., Phillips, G. J., Axinte, R., Fischer,H., Petäjä, T., Nölscher, A., Williams, J., Hens, K., Harder, H.,Martinez-Harder, M., Novelli, A., Kubistin, D., Bohn, B., andLelieveld, J.: Insights into HOx and ROx chemistry in the bo-real forest via measurement of peroxyacetic acid, peroxyaceticnitric anhydride (PAN) and hydrogen peroxide, Atmos. Chem.Phys., 18, 13457–13479, https://doi.org/10.5194/acp-18-13457-2018, 2018.

Dee, D. P., Uppala, S. M., Simmons, A. J., Berrisford, P., Poli,P., Kobayashi, S., Andrae, U., Balmaseda, M. A., Balsamo, G.,Bauer, P., Bechtold, P., Beljaars, A. C. M., van de Berg, L., Bid-lot, J., Bormann, N., Delsol, C., Dragani, R., Fuentes, M., Geer,A. J., Haimberger, L., Healy, S. B., Hersbach, H., Hólm, E. V.,Isaksen, L., Kållberg, P., Köhler, M., Matricardi, M., McNally,A. P., Monge-Sanz, B. M., Morcrette, J.-J., Park, B.-K., Peubey,C., de Rosnay, P., Tavolato, C., Thépaut, J.-N., and Vitart, F.: TheERA-Interim reanalysis: configuration and performance of thedata assimilation system, Q. J. Roy. Meteorol. Soc., 137, 553–597, https://doi.org/10.1002/qj.828, 2011.

Dibble, T. S.: Mechanism and dynamics of theCH2OH+O2 reaction, Chem. Phys. Lett., 355, 193–200,https://doi.org/10.1016/S0009-2614(02)00211-7, 2002.

Dillon, T. J., Pozzer, A., Vereecken, L., Crowley, J. N.,and Lelieveld, J.: Does acetone react with HO2 in theupper-troposphere?, Atmos. Chem. Phys., 12, 1339–1351,https://doi.org/10.5194/acp-12-1339-2012, 2012.

Douglas, K. M., Blitz, M. A., Feng, W. H., Heard, D. E.,Plane, J. M. C., Rashid, H., and Seakins, P. W.: Low tem-perature studies of the rate coefficients and branching ra-tios of reactive loss vs quenching for the reactions of(CH2)-C-1 with C2H6, C2H4, C2H2, Icarus, 321, 752–766,https://doi.org/10.1016/j.icarus.2018.12.027, 2019.

Dunning, T. H.: Gaussian-basis sets for use in correlated molecularcalculations. 1. the atoms boron through neon and hydrogen, J.Chem. Phys., 90, 1007–1023, https://doi.org/10.1063/1.456153,1989.

Eckart, C.: The penetration of a potential barrier by electrons, Phys.Rev., 35, 1303–1309, https://doi.org/10.1103/PhysRev.35.1303,1930.

Eger, P. G., Schuladen, J., Sobanski, N., Fischer, H., Karu, E.,Williams, J., Riva, M., Zha, Q., Ehn, M., Quéléver, L. L.J., Schallhart, S., Lelieveld, J., and Crowley, J. N.: Pyruvicacid in the boreal forest: gas-phase mixing ratios and impact

Atmos. Chem. Phys., 21, 14333–14349, 2021 https://doi.org/10.5194/acp-21-14333-2021

P. G. Eger et al.: Impact of pyruvic acid photolysis on CH3CHO and RO2 14347

on radical chemistry, Atmos. Chem. Phys., 20, 3697–3711,https://doi.org/10.5194/acp-20-3697-2020, 2020.

Eugene, A. J., Pillar, E. A., Colussi, A. J., and Guz-man, M. I.: Enhanced Acidity of Acetic and PyruvicAcids on the Surface of Water, Langmuir, 34, 9307–9313,https://doi.org/10.1021/acs.langmuir.8b01606, 2018.

Fischer, H., Axinte, R., Bozem, H., Crowley, J. N., Ernest, C.,Gilge, S., Hafermann, S., Harder, H., Hens, K., Janssen, R. H.H., Königstedt, R., Kubistin, D., Mallik, C., Martinez, M., Nov-elli, A., Parchatka, U., Plass-Dülmer, C., Pozzer, A., Regelin, E.,Reiffs, A., Schmidt, T., Schuladen, J., and Lelieveld, J.: Diur-nal variability, photochemical production and loss processes ofhydrogen peroxide in the boundary layer over Europe, Atmos.Chem. Phys., 19, 11953–11968, https://doi.org/10.5194/acp-19-11953-2019, 2019.

Fischer, L., Breitenlechner, M., Canaval, E., Scholz, W., Striednig,M., Graus, M., Karl, T. G., Petäjä, T., Kulmala, M., and Hansel,A.: First Eddy Covariance Flux Measurements of Semi VolatileOrganic Compounds with the PTR3-TOF-MS, Atmos. Meas.Tech. Discuss. [preprint], https://doi.org/10.5194/amt-2021-117,in review, 2021.

Frisch, M. J., Trucks, G. W., Schlegel, H. B., Scuseria, G. E.,Robb, M. A., Cheeseman, J. R., Scalmani, G., Barone, V., Men-nucci, B., Petersson, G. A., Nakatsuji, H., Li, X., Caricato, M.,Marenich, A. V., Bloino, J., Janesko, B. G., Gomperts, R., Men-nucci, B., Hratchian, H. P., Ortiz, J. V., Izmaylov, A. F., Sonnen-berg, J. L., Williams-Young, D., Ding, F., Lipparini, F., Egidi,F., Goings, J., Peng, B., Petrone, A., Henderson, T., Ranasinghe,D., Zakrzewski, V. G., Gao, J., Rega, N., Zheng, G., Liang,W., Hada, M., Eara, M., Toyota, K., Fukuda, R., Hasegawa,J., Ishida, M., Nakajima, T., Honda, Y., Kitao, O., Nakai, H.,Vreven, T., Throssell, K., Mntgomery, Jr., J. A., Peralta, J. E.,Ogliaro, F., Bearpark, M. J., Heyd, J. J., Brothers, E. N., Kudin,K. N., Staroverov, V. N., Keith, T. A., Kobayashi, R., Normand,J., Raghavachari, K., Rendell, A. P., Burant, J. C., Iyengar, S.S., Tomasi, J., Cossi, M., Millam, J. M., Klene, M., Adamo,C., Cammi, R., Ochterski, J. W., Martin, R. L., Morokuma, K.,Farkas, O., Foresman, J. B., and Fox, D. J.: Gaussian 16, Revi-sion B.01, Gaussian Inc., Wallington, CT, 2016.

Gannon, K. L., Blitz, M. A., Liang, C. H., Pilling, M. J., Seakins,P. W., and Glowacki, D. R.: Temperature Dependent Kinetics(195–798 K) and H Atom Yields (298–498 K) from Reactionsof (CH2)-C-1 with Acetylene, Ethene, and Propene, J. Phys.Chem. A, 114, 9413–9424, https://doi.org/10.1021/jp102276j,2010.

Gaona-Colman, E., Blanco, M. B., Barnes, I., Wiesen, P., andTeruel, M. A.: OH- and O-3-initiated atmospheric degrada-tion of camphene: temperature dependent rate coefficients,product yields and mechanisms, RSC Adv., 7, 2733–2744,https://doi.org/10.1039/c6ra26656h, 2017.

Goulay, F., Trevitt, A. J., Meloni, G., Selby, T. M., Osborn, D. L.,Taatjes, C. A., Vereecken, L., and Leone, S. R.: Cyclic VersusLinear Isomers Produced by Reaction of the Methylidyne Rad-ical (CH) with Small Unsaturated Hydrocarbons, J. Am. Chem.Soc., 131, 993–1005, https://doi.org/10.1021/ja804200v, 2009.

Grimme, S., Ehrlich, S., and Goerigk, L.: Effect of theDamping Function in Dispersion Corrected DensityFunctional Theory, J. Comput. Chem., 32, 1456–1465,https://doi.org/10.1002/jcc.21759, 2011.

Grosjean, D.: Atmospheric reactions of pyruvic acid, At-mos. Environ., 17, 2379–2382, https://doi.org/10.1016/0004-6981(83)90242-1, 1983.

Grosjean, D.: Atmospheric reactions of ortho cresol: gas phaseand aerosol products, Atmos. Environ., 18, 1641–1652,https://doi.org/10.1016/0004-6981(84)90386-X, 1984.

Grosjean, D., Williams, E. L., and Grosjean, E.: Atmospheric chem-istry of isoprene and of its carbonyl products, Environ. Sci. Tech-nol., 27, 830–840, https://doi.org/10.1021/es00042a004, 1993.

Guenther, A. B., Zimmerman, P. R., Harley, P. C., Monson, R. K.,and Fall, R.: Isoprene and monoterpene emission rate variability- model evaluations and sensitivity analyses, J. Geophys. Res.-Atmos., 98, 12609–12617, https://doi.org/10.1029/93jd00527,1993.

Hakola, H., Hellén, H., and Laurila, T.: Ten years of lighthydrocarbons (C2–C6) concentration measurements in back-ground air in Finland, Atmos. Environ., 40, 3621–3630,https://doi.org/10.1016/j.atmosenv.2005.08.019, 2006.

Hallquist, M., Wenger, J. C., Baltensperger, U., Rudich, Y., Simp-son, D., Claeys, M., Dommen, J., Donahue, N. M., George,C., Goldstein, A. H., Hamilton, J. F., Herrmann, H., Hoff-mann, T., Iinuma, Y., Jang, M., Jenkin, M. E., Jimenez, J. L.,Kiendler-Scharr, A., Maenhaut, W., McFiggans, G., Mentel, Th.F., Monod, A., Prévôt, A. S. H., Seinfeld, J. H., Surratt, J. D.,Szmigielski, R., and Wildt, J.: The formation, properties and im-pact of secondary organic aerosol: current and emerging issues,Atmos. Chem. Phys., 9, 5155–5236, https://doi.org/10.5194/acp-9-5155-2009, 2009.

Hari, P. and Kulmala, M.: Station for Measuring Ecosystem–Atmosphere Relations (SMEAR II), Boreal Environ. Res., 10,315–322, https://doi.org/10.1007/978-94-007-5603-8_9, 2005.

Helas, G., Bingemer, H., and Andreae, M. O.: Organic acids overequatorial Africa: Results from DECAFE 88, J. Geophys. Res.-Atmos., 97, 6187–6193, https://doi.org/10.1029/91jd01438,1992.

Hellén, H., Kouznetsov, R., Anttila, P., and Hakola, H.: Increasinginfluence of easterly air masses on NMHC concentrations at thePallas-Sodankylä GAW station, Boreal Environ. Res., 20, 542–552, 2015.

Hellén, H., Praplan, A. P., Tykkä, T., Ylivinkka, I., Vakkari,V., Bäck, J., Petäjä, T., Kulmala, M., and Hakola, H.: Long-term measurements of volatile organic compounds highlightthe importance of sesquiterpenes for the atmospheric chem-istry of a boreal forest, Atmos. Chem. Phys., 18, 13839–13863,https://doi.org/10.5194/acp-18-13839-2018, 2018.

Hermans, I., Nguyen, T. L., Jacobs, P. A., and Peeters, J.:Tropopause chemistry revisited: HO2-initiated oxidation as anefficient acetone sink, J. Am. Chem. Soc., 126, 9908–9909,https://doi.org/10.1021/ja0467317, 2004.

Hermans, I., Muller, J. F., Nguyen, T. L., Jacobs, P. A., andPeeters, J.: Kinetics of alpha-hydroxy-alkylperoxyl radicals inoxidation processes. HO2 center dot-initiated oxidation of ke-tones/aldehydes near the tropopause, J. Phys. Chem. A, 109,4303–4311, 2005.

IUPAC Task Group on Atmospheric Chemical Kinetic Data Evalu-ation (Ammann, M., Cox, R. A., Crowley, J. N., Herrmann, H.,Jenkin, M. E., McNeill, V. F., Mellouki, A., Rossi, M. J., Troe, J.,and Wallington, T. J.): Evaluated kinetic and photochemical data

https://doi.org/10.5194/acp-21-14333-2021 Atmos. Chem. Phys., 21, 14333–14349, 2021

14348 P. G. Eger et al.: Impact of pyruvic acid photolysis on CH3CHO and RO2

for atmospheric chemistry, available at: http://iupac.pole-ether.fr/index.html, last access: June 2021.

Jacob, D. J. and Wofsy, S. C.: Photochemistry of biogenic emissionsover the Amazon forest, J. Geophys. Res.-Atmos., 93, 1477–1486, https://doi.org/10.1029/JD093iD02p01477, 1988.

Jardine, K., Abrell, L., Kurc, S. A., Huxman, T., Ortega, J., andGuenther, A.: Volatile organic compound emissions from Lar-rea tridentata (creosotebush), Atmos. Chem. Phys., 10, 12191–12206, https://doi.org/10.5194/acp-10-12191-2010, 2010a.

Jardine, K. J., Sommer, E. D., Saleska, S. R., Huxman, T. E., Harley,P. C., and Abrell, L.: Gas Phase Measurements of Pyruvic Acidand Its Volatile Metabolites, Environ. Sci. Technol., 44, 2454–2460, https://doi.org/10.1021/es903544p, 2010b.

Jasper, A. W., Klippenstein, S. J., and Harding, L. B.: Secondary Ki-netics of Methanol Decomposition: Theoretical Rate Coefficientsfor 3CH2

+OH, 3CH2+ 3CH2, and 3CH2+CH3, J. Phys.Chem. A, 111, 8699–8707, https://doi.org/10.1021/jp0736950,2007.

Johnston, H. S. and Heicklen, J.: Tunnelling corrections for un-symmetrical eckart potential energy barriers, J. Phys. Chem., 66,532–533, https://doi.org/10.1021/j100809a040, 1962.

Kanakidou, M., Seinfeld, J. H., Pandis, S. N., Barnes, I., Dentener,F. J., Facchini, M. C., Van Dingenen, R., Ervens, B., Nenes, A.,Nielsen, C. J., Swietlicki, E., Putaud, J. P., Balkanski, Y., Fuzzi,S., Horth, J., Moortgat, G. K., Winterhalter, R., Myhre, C. E.L., Tsigaridis, K., Vignati, E., Stephanou, E. G., and Wilson, J.:Organic aerosol and global climate modelling: a review, Atmos.Chem. Phys., 5, 1053–1123, https://doi.org/10.5194/acp-5-1053-2005, 2005.

Liebmann, J., Karu, E., Sobanski, N., Schuladen, J., Ehn, M.,Schallhart, S., Quéléver, L., Hellen, H., Hakola, H., Hoffmann,T., Williams, J., Fischer, H., Lelieveld, J., and Crowley, J. N.: Di-rect measurement of NO3 radical reactivity in a boreal forest, At-mos. Chem. Phys., 18, 3799–3815, https://doi.org/10.5194/acp-18-3799-2018, 2018a.

Liebmann, J. M., Muller, J. B. A., Kubistin, D., Claude, A., Holla,R., Plass-Dülmer, C., Lelieveld, J., and Crowley, J. N.: Directmeasurements of NO3 reactivity in and above the boundary layerof a mountaintop site: identification of reactive trace gases andcomparison with OH reactivity, Atmos. Chem. Phys., 18, 12045–12059, https://doi.org/10.5194/acp-18-12045-2018, 2018b.

Liebmann, J., Sobanski, N., Schuladen, J., Karu, E., Hellén, H.,Hakola, H., Zha, Q., Ehn, M., Riva, M., Heikkinen, L., Williams,J., Fischer, H., Lelieveld, J., and Crowley, J. N.: Alkyl ni-trates in the boreal forest: formation via the NO3-, OH- andO3-induced oxidation of biogenic volatile organic compoundsand ambient lifetimes, Atmos. Chem. Phys., 19, 10391–10403,https://doi.org/10.5194/acp-19-10391-2019, 2019.

Mellouki, A. and Mu, Y.: On the atmospheric degradation of pyru-vic acid in the gas phase, J. Photoch. Photobio. A, 157, 295–300,https://doi.org/10.1016/S1010-6030(03)00070-4, 2003.

Millet, D. B., Guenther, A., Siegel, D. A., Nelson, N. B., Singh,H. B., de Gouw, J. A., Warneke, C., Williams, J., Eerdekens,G., Sinha, V., Karl, T., Flocke, F., Apel, E., Riemer, D. D.,Palmer, P. I., and Barkley, M.: Global atmospheric budget ofacetaldehyde: 3-D model analysis and constraints from in-situand satellite observations, Atmos. Chem. Phys., 10, 3405–3425,https://doi.org/10.5194/acp-10-3405-2010, 2010.

Millet, D. B., Baasandorj, M., Farmer, D. K., Thornton, J. A., Bau-mann, K., Brophy, P., Chaliyakunnel, S., de Gouw, J. A., Graus,M., Hu, L., Koss, A., Lee, B. H., Lopez-Hilfiker, F. D., Neu-man, J. A., Paulot, F., Peischl, J., Pollack, I. B., Ryerson, T. B.,Warneke, C., Williams, B. J., and Xu, J.: A large and ubiqui-tous source of atmospheric formic acid, Atmos. Chem. Phys., 15,6283–6304, https://doi.org/10.5194/acp-15-6283-2015, 2015.

Noodleman, L.: Valence bond description of anti-ferromagneticcoupling in transition-metal dimers, J. Chem. Phys., 74, 5737–5743, 1981.

Olivella, S., Bofill, J. M., and Solé, A.: Ab Initio Calculationson the Mechanism of the Oxidation of the HydroxymethylRadical by Molecular Oxygen in the Gas Phase: A SignificantReaction for Environmental Science, Chemistry, 7, 3377–3386,https://doi.org/10.1002/1521-3765(20010803)7:15<3377::AID-CHEM3377>3.0.CO;2-8, 2001.

Paulot, F., Crounse, J. D., Kjaergaard, H. G., Kroll, J. H., Sein-feld, J. H., and Wennberg, P. O.: Isoprene photooxidation: newinsights into the production of acids and organic nitrates, Atmos.Chem. Phys., 9, 1479–1501, https://doi.org/10.5194/acp-9-1479-2009, 2009.

Perring, A. E., Pusede, S. E., and Cohen, R. C.: An observationalperspective on the atmospheric impacts of alkyl and multifunc-tional nitrates on ozone and secondary organic aerosol, Chem.Rev., 113, 5848–5870, https://doi.org/10.1021/cr300520x, 2013.

Petäjä, T., Mauldin, III, R. L., Kosciuch, E., McGrath, J., Nieminen,T., Paasonen, P., Boy, M., Adamov, A., Kotiaho, T., and Kulmala,M.: Sulfuric acid and OH concentrations in a boreal forest site,Atmos. Chem. Phys., 9, 7435–7448, https://doi.org/10.5194/acp-9-7435-2009, 2009.

Praplan, A. P., Hegyi-Gaeggeler, K., Barmet, P., Pfaffenberger,L., Dommen, J., and Baltensperger, U.: Online measurementsof water-soluble organic acids in the gas and aerosol phasefrom the photooxidation of 1,3,5-trimethylbenzene, Atmos.Chem. Phys., 14, 8665–8677, https://doi.org/10.5194/acp-14-8665-2014, 2014.

Purvis, G. D. and Bartlett, R. J.: A full coupled-cluster singles anddoubles model – the inclusion of disconnected triples, J. Chem.Phys., 76, 1910–1918, https://doi.org/10.1063/1.443164, 1982.

Raber, W. H. and Moortgat, G. K.: Photooxidation of selected car-bonyl compounds in air: methyl ethyl ketone, methyl vinyl ke-tone, methacrolein and methylglyoxal, in: Progress and problemsin atmospheric chemistry, edited by: Barker, J. R., World Scien-tific Publishing, Singapore, 318–373, 1995.

Reed Harris, A. E., Doussin, J. F., Carpenter, B. K., and Vaida, V.:Gas-Phase Photolysis of Pyruvic Acid: The Effect of Pressureon Reaction Rates and Products, J. Phys. Chem. A, 120, 10123–10133, https://doi.org/10.1021/acs.jpca.6b09058, 2016.

Reed Harris, A. E., Cazaunau, M., Gratien, A., Pangui, E., Doussin,J.-F., and Vaida, V.: Atmospheric Simulation Chamber Studiesof the Gas-Phase Photolysis of Pyruvic Acid, J. Phys. Chem.A, 121, 8348–8358, https://doi.org/10.1021/acs.jpca.7b05139,2017a.

Reed Harris, A. E., Cazaunau, M., Gratien, A., Pangui, E., Doussin,J. F., and Vaida, V.: Atmospheric Simulation Chamber Studiesof the Gas-Phase Photolysis of Pyruvic Acid, J. Phys. Chem.A, 121, 8348–8358, https://doi.org/10.1021/acs.jpca.7b05139,2017b.

Atmos. Chem. Phys., 21, 14333–14349, 2021 https://doi.org/10.5194/acp-21-14333-2021

P. G. Eger et al.: Impact of pyruvic acid photolysis on CH3CHO and RO2 14349

Rinne, J., Hakola, H., Laurila, T., and Rannik, Ü.: Canopy scalemonoterpene emissions of Pinus sylvestris dominated forests,Atmos. Environ., 34, 1099–1107, https://doi.org/10.1016/S1352-2310(99)00335-0, 2000.

Roberts, J. M., Flocke, F., Chen, G., de Gouw, J., Holloway, J.S., Hübler, G., Neuman, J. A., Nicks Jr., D. K., Nowak, J.B., Parrish, D. D., Ryerson, T. B., Sueper, D. T., Warneke,C., and Fehsenfeld, F. C.: Measurement of peroxycarboxylicnitric anhydrides (PANs) during the ITCT 2K2 aircraft in-tensive experiment, J. Geophys. Res.-Atmos., 109, D23S21,https://doi.org/10.1029/2004jd004960, 2004.

Roiger, A., Aufmhoff, H., Stock, P., Arnold, F., and Schlager, H.:An aircraft-borne chemical ionization – ion trap mass spec-trometer (CI-ITMS) for fast PAN and PPN measurements, At-mos. Meas. Tech., 4, 173–188, https://doi.org/10.5194/amt-4-173-2011, 2011.

Samanta, B. R., Fernando, R., Rösch, D., Reisler, H., and Osborn,D. L.: Primary photodissociation mechanisms of pyruvic acid onS1: observation of methylhydroxycarbene and its chemical reac-tion in the gas phase, Phys. Chem. Chem. Phys., 23, 4107–4119,https://doi.org/10.1039/D0CP06424F, 2021.

Sander, R., Baumgaertner, A., Cabrera-Perez, D., Frank, F., Gro-mov, S., Grooß, J.-U., Harder, H., Huijnen, V., Jöckel, P., Kary-dis, V. A., Niemeyer, K. E., Pozzer, A., Riede, H., Schultz,M. G., Taraborrelli, D., and Tauer, S.: The community atmo-spheric chemistry box model CAABA/MECCA-4.0, Geosci.Model Dev., 12, 1365–1385, https://doi.org/10.5194/gmd-12-1365-2019, 2019.

Schreiner, P. R., Reisenauer, H. P., Ley, D., Gerbig, D., Wu,C. H., and Allen, W. D.: Methylhydroxycarbene: Tunnel-ing control of a chemical reaction, Science, 332, 1300–1303,https://doi.org/10.1126/science.1203761, 2011.

Shepson, P. B., Bottenheim, J. W., Hastie, D. R., and Venka-tram, A.: Determination of the relative ozone and PAN depo-sition velocities at night, Geophys. Res. Lett., 19, 1121–1124,https://doi.org/10.1029/92gl01118, 1992a.

Shepson, P. B., Hastie, D. R., So, K. W., and Schiff, H. I.:Relationships between PAN, PPN and O3 at urban and ru-ral sites in ontario, Atmos. Environ. A-Gen., 26, 1259–1270,https://doi.org/10.1016/0960-1686(92)90387-z, 1992b.

Sobanski, N., Schuladen, J., Schuster, G., Lelieveld, J., andCrowley, J. N.: A five-channel cavity ring-down spectrometerfor the detection of NO2, NO3, N2O5, total peroxy nitratesand total alkyl nitrates, Atmos. Meas. Tech., 9, 5103–5118,https://doi.org/10.5194/amt-9-5103-2016, 2016.

Stefan, M. I. and Bolton, J. R.: Reinvestigation of the ace-tone degradation mechanism in dilute aqueous solution bythe UV/H2O2 process, Environ. Sci. Technol., 33, 870–873,https://doi.org/10.1021/es9808548, 1999.

Talbot, R., Andreae, M., Berresheim, H., Jacob, D. J., and Beecher,K.: Sources and sinks of formic, acetic, and pyruvic acids overCentral Amazonia: 2. Wet season, J. Geophys. Res.-Atmos., 95,16799–16811, https://doi.org/10.1029/JD095iD10p16799, 1990.

Vereecken, L. and Peeters, J.: The 1,5-H-shift in 1-butoxy: A casestudy in the rigorous implementation of transition state theoryfor a multirotamer system, J. Chem. Phys., 119, 5159–5170,https://doi.org/10.1063/1.1597479, 2003.

Vereecken, L., Pierloot, K., and Peeters, J.: B3LYP-DFTcharacterization of the potential energy surface of theCH(X25)+C2H2 reaction, J. Chem. Phys., 108, 1068–1080,https://doi.org/10.1063/1.475345, 1998.

Vesley, G. F. and Leermakers, P. A.: Photochemistry ofalpha-keto acids+ alpha-keto esters. 3. Photolysis of pyru-vic acid in vapor phase, J. Phys. Chem., 68, 2364–2366,https://doi.org/10.1021/j100790a507, 1964.

Walker, D.: Pyruvate carboxylation and plant metabolism,Biol. Rev., 37, 215–254, https://doi.org/10.1111/j.1469-185X.1962.tb01611.x, 1962.

Wang, N., Edtbauer, A., Stönner, C., Pozzer, A., Bourtsoukidis,E., Ernle, L., Dienhart, D., Hottmann, B., Fischer, H., Schu-laden, J., Crowley, J. N., Paris, J.-D., Lelieveld, J., and Williams,J.: Measurements of carbonyl compounds around the ArabianPeninsula: overview and model comparison, Atmos. Chem.Phys., 20, 10807–10829, https://doi.org/10.5194/acp-20-10807-2020, 2020.

Wang, S. Y., Hornbrook, R. S., Hills, A., Emmons, L. K., Tilmes,S., Lamarque, J. F., Jimenez, J. L., Campuzano-Jost, P., Nault, E.A., Crounse, J. D., Wennberg, P. O., Kim, M., Allen, H., Ryerson,T. B., Thompson, C. R., Peischl, J., Moore, F., Nance, D., Hall,B., Elkins, J., Tanner, D., Huey, L. G., Hall, S. R., Ullmann, K.,Orlando, J. J., Tyndall, G. S., Flocke, F. M., Ray, E., Hanisco, T.F., Wolfe, G. M., St Clair, J., Commane, R., Daube, B., Barletta,B., Blake, D. R., Weinzierl, B., Dollner, M., Conley, A., Vitt, F.,Wofsy, S. C., Riemer, D. D., and Apel, E. C.: Atmospheric ac-etaldehyde: importance of air-sea exchange and a missing sourcein the remote troposphere, Geophys. Res. Lett., 46, 5601–5613,https://doi.org/10.1029/2019gl082034, 2019.

Williams, J., Crowley, J., Fischer, H., Harder, H., Martinez, M.,Petäjä, T., Rinne, J., Bäck, J., Boy, M., Dal Maso, M., Hakala,J., Kajos, M., Keronen, P., Rantala, P., Aalto, J., Aaltonen,H., Paatero, J., Vesala, T., Hakola, H., Levula, J., Pohja, T.,Herrmann, F., Auld, J., Mesarchaki, E., Song, W., Yassaa, N.,Nölscher, A., Johnson, A. M., Custer, T., Sinha, V., Thieser,J., Pouvesle, N., Taraborrelli, D., Tang, M. J., Bozem, H.,Hosaynali-Beygi, Z., Axinte, R., Oswald, R., Novelli, A., Ku-bistin, D., Hens, K., Javed, U., Trawny, K., Breitenberger, C.,Hidalgo, P. J., Ebben, C. J., Geiger, F. M., Corrigan, A. L.,Russell, L. M., Ouwersloot, H. G., Vilà-Guerau de Arellano,J., Ganzeveld, L., Vogel, A., Beck, M., Bayerle, A., Kampf,C. J., Bertelmann, M., Köllner, F., Hoffmann, T., Valverde, J.,González, D., Riekkola, M.-L., Kulmala, M., and Lelieveld,J.: The summertime Boreal forest field measurement intensive(HUMPPA-COPEC-2010): an overview of meteorological andchemical influences, Atmos. Chem. Phys., 11, 10599–10618,https://doi.org/10.5194/acp-11-10599-2011, 2011.

Yamamoto, S. and Back, R. A.: The photolysis and thermal-decomposition of pyruvic-acid in the gas-phase, Can. J. Chem.,63, 549–554, https://doi.org/10.1139/v85-089, 1985.

Zhao, Y. and Truhlar, D. G.: The M06 suite of density functionalsfor main group thermochemistry, thermochemical kinetics, non-covalent interactions, excited states, and transition elements: twonew functionals and systematic testing of four M06-class func-tionals and 12 other functionals, Theor. Chem. Account., 120,215–241, https://doi.org/10.1007/s00214-007-0310-x, 2008.

https://doi.org/10.5194/acp-21-14333-2021 Atmos. Chem. Phys., 21, 14333–14349, 2021